1. Chapter 9
Lecture 4
Section: 9.4

2. We have studied inferences about the means of two independent populations. We defined two samples to be independent if the sample values selected from one population are not related to, paired or matched with the sample values from the other population. We will now study dependent samples, which we refer to as matched pairs.
With matched pairs, there is some relationship so that each value in one sample is paired with a corresponding value in the other sample.
Examples:
1. When conducting an experiment to test the effectiveness of a low-fat diet, the weight of each subject is measured once before the diet and once after the diet.
2. The accuracy of reported weights is analyzed with a sample of people when, for each person, the reported weight is recorded and the actual weight is measured.

3. Assumptions:
1. The sample data consist of matched pairs.
2. The samples are simple random samples.
3. Either or both of these conditions is satisfied:
i. The number of matched pairs of sample data is large, n > 30
ii. The pairs of values have differences that are from a population
having a distribution that is approximately normal.
Notation or Matched Pairs:

4. Confidence Intervals Estimate for Matched Pairs:
df = n – 1
Test Statistic for Matched Pairs:
Where H0: μd=0

5. 3. 6 people were randomly selected and asked their hourly wage
3. 6 people were randomly selected and asked their hourly wage. Then they were asked to show proof of their hourly wage. Suppose the samples are taken from a normal population, given below.
Test the claim that there is a difference between reported and verified hourly wages at α =
Reported:
Verified:
4. Construct a 99% confidence interval for the difference in the two populations for problem #1.

6. 5. A random sample of 10 patients is given some new brand of prescription drug to reduce blood pressure. The samples are taken from a normal population. Test the claim that the new prescription drug is reducing the blood pressure significantly at a 0.05 level.
Before:
After:
6. Construct a 90% confidence interval for the difference in the two populations for problem #5.

7. 7. Listed below are heights, in inches, of females statistic students
7. Listed below are heights, in inches, of females statistic students. Is there sufficient evidence to support the claim that female students exaggerate by reporting heights that are greater than their actual measured heights? Use a 0.10 level of significance.
Reported:
Actual:
8. Construct a 90% confidence interval for the difference in the two populations for problem #7.

8. Minitab Output:
Paired T-Test and CI: Reported, Measured
Paired T for Reported - Measured
N Mean StDev SE Mean
Reported
Measured
Difference
90% lower bound for mean difference:
T-Test of mean difference = 0 (vs > 0): T-Value = P-Value = 0.026
Paired T-Test and CI: Reported, Measured
Paired T for Reported - Measured
N Mean StDev SE Mean
Reported
Measured
Difference
90% CI for mean difference: ( , )
T-Test of mean difference = 0 (vs not = 0): T-Value = P-Value = 0.052

9. 9. A listening test was administered to English teachers before and after an institute designed to improve English listening skills.
The maximum possible score on the test was 36:
Sub
Pre
Post
Determine if the institute improved listening skills at the 5% significance level.
10. Determine if a 99% confidence interval will give the same conclusion as the hypothesis test above.